Convolutional Neural Networks

Project: Write an Algorithm for a Dog Identification App


In this notebook, some template code has already been provided for you, and you will need to implement additional functionality to successfully complete this project. You will not need to modify the included code beyond what is requested. Sections that begin with '(IMPLEMENTATION)' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section, and the specifics of the implementation are marked in the code block with a 'TODO' statement. Please be sure to read the instructions carefully!

Note: Once you have completed all of the code implementations, you need to finalize your work by exporting the Jupyter Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.

In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.

Note: Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. Markdown cells can be edited by double-clicking the cell to enter edit mode.

The rubric contains optional "Stand Out Suggestions" for enhancing the project beyond the minimum requirements. If you decide to pursue the "Stand Out Suggestions", you should include the code in this Jupyter notebook.


Why We're Here

In this notebook, you will make the first steps towards developing an algorithm that could be used as part of a mobile or web app. At the end of this project, your code will accept any user-supplied image as input. If a dog is detected in the image, it will provide an estimate of the dog's breed. If a human is detected, it will provide an estimate of the dog breed that is most resembling. The image below displays potential sample output of your finished project (... but we expect that each student's algorithm will behave differently!).

Sample Dog Output

In this real-world setting, you will need to piece together a series of models to perform different tasks; for instance, the algorithm that detects humans in an image will be different from the CNN that infers dog breed. There are many points of possible failure, and no perfect algorithm exists. Your imperfect solution will nonetheless create a fun user experience!

The Road Ahead

We break the notebook into separate steps. Feel free to use the links below to navigate the notebook.

  • Step 0: Import Datasets
  • Step 1: Detect Humans
  • Step 2: Detect Dogs
  • Step 3: Create a CNN to Classify Dog Breeds (from Scratch)
  • Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)
  • Step 5: Write your Algorithm
  • Step 6: Test Your Algorithm

Step 0: Import Datasets

Make sure that you've downloaded the required human and dog datasets:

Note: if you are using the Udacity workspace, you DO NOT need to re-download these - they can be found in the /data folder as noted in the cell below.

  • Download the dog dataset. Unzip the folder and place it in this project's home directory, at the location /dog_images.

  • Download the human dataset. Unzip the folder and place it in the home directory, at location /lfw.

Note: If you are using a Windows machine, you are encouraged to use 7zip to extract the folder.

In the code cell below, we save the file paths for both the human (LFW) dataset and dog dataset in the numpy arrays human_files and dog_files.

In [1]:
import numpy as np
from glob import glob

# load filenames for human and dog images
human_files = np.array(glob("/data/lfw/*/*"))
dog_files = np.array(glob("/data/dog_images/*/*/*"))

# print number of images in each dataset
print('There are %d total human images.' % len(human_files))
print('There are %d total dog images.' % len(dog_files))
There are 13233 total human images.
There are 8351 total dog images.

Step 1: Detect Humans

In this section, we use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images.

OpenCV provides many pre-trained face detectors, stored as XML files on github. We have downloaded one of these detectors and stored it in the haarcascades directory. In the next code cell, we demonstrate how to use this detector to find human faces in a sample image.

In [2]:
import cv2                
import matplotlib.pyplot as plt                        
%matplotlib inline                               

# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')

def human_face(images_path):
    
    # load color (BGR) image
    img = cv2.imread(images_path)
    # convert BGR image to grayscale
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

    # find faces in image
    faces = face_cascade.detectMultiScale(gray)

    # print number of faces detected in the image
    print('Number of faces detected:', len(faces))

    # get bounding box for each detected face
    for (x,y,w,h) in faces:
        # add bounding box to color image
        cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
    
    # convert BGR image to RGB for plotting
    cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

    # display the image, along with bounding box
    plt.imshow(cv_rgb)
    plt.show()
In [3]:
human_face(human_files[0])
Number of faces detected: 1

Before using any of the face detectors, it is standard procedure to convert the images to grayscale. The detectMultiScale function executes the classifier stored in face_cascade and takes the grayscale image as a parameter.

In the above code, faces is a numpy array of detected faces, where each row corresponds to a detected face. Each detected face is a 1D array with four entries that specifies the bounding box of the detected face. The first two entries in the array (extracted in the above code as x and y) specify the horizontal and vertical positions of the top left corner of the bounding box. The last two entries in the array (extracted here as w and h) specify the width and height of the box.

Write a Human Face Detector

We can use this procedure to write a function that returns True if a human face is detected in an image and False otherwise. This function, aptly named face_detector, takes a string-valued file path to an image as input and appears in the code block below.

In [4]:
# returns "True" if face is detected in image stored at img_path
def face_detector(img_path):
    img = cv2.imread(img_path)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray)
    return len(faces) > 0

(IMPLEMENTATION) Assess the Human Face Detector

Question 1: Use the code cell below to test the performance of the face_detector function.

  • What percentage of the first 100 images in human_files have a detected human face?
  • What percentage of the first 100 images in dog_files have a detected human face?

Ideally, we would like 100% of human images with a detected face and 0% of dog images with a detected face. You will see that our algorithm falls short of this goal, but still gives acceptable performance. We extract the file paths for the first 100 images from each of the datasets and store them in the numpy arrays human_files_short and dog_files_short.

Answer: Based on the printed statment below, the detector deceted some human images whithin the dog images. Therefore, I used the implemented code before to print the detected human faces. Clearly, we can see that most of the images don't have any human faces, except for four.

In [5]:
from tqdm import tqdm

human_files_short = human_files[:100]
dog_files_short = dog_files[:100]

#-#-# Do NOT modify the code above this line. #-#-#

## TODO: Test the performance of the face_detector algorithm 
## on the images in human_files_short and dog_files_short.

def human_images_counter(file_path):
    
    # counter to count number of human images
    c = []
    
    for i in file_path:
        
        if(face_detector(i)):
            c.append(i)

    return c
In [6]:
paths1 = human_images_counter(human_files_short)
print('The first 100 file paths contains {}% human images'.format(len(paths1)/len(human_files_short)*100))

paths2 = human_images_counter(dog_files_short)
print('The second 100 file paths contains {}% human images'.format(len(paths2)/len(dog_files_short)*100))
The first 100 file paths contains 98.0% human images
The second 100 file paths contains 17.0% human images
In [7]:
for i in paths2:
    human_face(i)
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 3
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 2
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1
Number of faces detected: 1

We suggest the face detector from OpenCV as a potential way to detect human images in your algorithm, but you are free to explore other approaches, especially approaches that make use of deep learning :). Please use the code cell below to design and test your own face detection algorithm. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [ ]:
### (Optional) 
### TODO: Test performance of anotherface detection algorithm.
### Feel free to use as many code cells as needed.

Step 2: Detect Dogs

In this section, we use a pre-trained model to detect dogs in images.

Obtain Pre-trained VGG-16 Model

The code cell below downloads the VGG-16 model, along with weights that have been trained on ImageNet, a very large, very popular dataset used for image classification and other vision tasks. ImageNet contains over 10 million URLs, each linking to an image containing an object from one of 1000 categories.

In [8]:
import torch
import torchvision.models as models

# define VGG16 model
VGG16 = models.vgg16(pretrained=True)

# check if CUDA is available
use_cuda = torch.cuda.is_available()

# move model to GPU if CUDA is available
if use_cuda:
    VGG16 = VGG16.cuda()
Downloading: "https://download.pytorch.org/models/vgg16-397923af.pth" to /root/.torch/models/vgg16-397923af.pth
100%|██████████| 553433881/553433881 [00:06<00:00, 88943683.08it/s] 

Given an image, this pre-trained VGG-16 model returns a prediction (derived from the 1000 possible categories in ImageNet) for the object that is contained in the image.

(IMPLEMENTATION) Making Predictions with a Pre-trained Model

In the next code cell, you will write a function that accepts a path to an image (such as 'dogImages/train/001.Affenpinscher/Affenpinscher_00001.jpg') as input and returns the index corresponding to the ImageNet class that is predicted by the pre-trained VGG-16 model. The output should always be an integer between 0 and 999, inclusive.

Before writing the function, make sure that you take the time to learn how to appropriately pre-process tensors for pre-trained models in the PyTorch documentation.

In [65]:
from PIL import Image
import torchvision.transforms as transforms

# Set PIL to be tolerant of image files that are truncated.
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True

def VGG16_predict(img_path):
    '''
    Use pre-trained VGG-16 model to obtain index corresponding to 
    predicted ImageNet class for image at specified path
    
    Args:
        img_path: path to an image
        
    Returns:
        Index corresponding to VGG-16 model's prediction
    '''
    
    ## TODO: Complete the function.
    ## Load and pre-process an image from the given img_path
    ## Return the *index* of the predicted class for that image
    
    
    #load the image 
    img = Image.open(img_path).convert('RGB')
    
    
    #pre-processing parameters (the same parameters values used in Pytorch documentation)
    transform = transforms.Compose([transforms.Resize(256),
                                   transforms.CenterCrop(224),
                                   transforms.ToTensor(),
                                   transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                         std=[0.229, 0.224, 0.225])])
    
    #transform the image 
    img = transform(img)
    img = img.unsqueeze(0)
    
    if use_cuda:
        img = img.cuda()
        
    #predection 
    pre = VGG16(img)
    
    return torch.argmax(pre).item() # predicted class index

(IMPLEMENTATION) Write a Dog Detector

While looking at the dictionary, you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from 'Chihuahua' to 'Mexican hairless'. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained VGG-16 model, we need only check if the pre-trained model predicts an index between 151 and 268 (inclusive).

Use these ideas to complete the dog_detector function below, which returns True if a dog is detected in an image (and False if not).

In [10]:
### returns "True" if a dog is detected in the image stored at img_path
def dog_detector(img_path):
    ## TODO: Complete the function.
    if VGG16_predict(img_path) >= 151 and VGG16_predict(img_path) <= 268:
        return True
    else:
        return False

(IMPLEMENTATION) Assess the Dog Detector

Question 2: Use the code cell below to test the performance of your dog_detector function.

  • What percentage of the images in human_files_short have a detected dog?
  • What percentage of the images in dog_files_short have a detected dog?

Answer:

In [11]:
### TODO: Test the performance of the dog_detector function
### on the images in human_files_short and dog_files_short.
def dog_image_counter(paths):

    counter = 0
    
    for i in paths:
        
        if (dog_detector(i)):
            counter += 1
    
    return counter/len(paths)*100
In [12]:
print('The first 100 file paths contains {}% dogs images'.format(dog_image_counter(human_files_short)))
print('The second 100 file paths contains {}% dogs images'.format(dog_image_counter(dog_files_short)))
The first 100 file paths contains 0.0% dogs images
The second 100 file paths contains 100.0% dogs images

We suggest VGG-16 as a potential network to detect dog images in your algorithm, but you are free to explore other pre-trained networks (such as Inception-v3, ResNet-50, etc). Please use the code cell below to test other pre-trained PyTorch models. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [ ]:
### (Optional) 
### TODO: Report the performance of another pre-trained network.
### Feel free to use as many code cells as needed.

Step 3: Create a CNN to Classify Dog Breeds (from Scratch)

Now that we have functions for detecting humans and dogs in images, we need a way to predict breed from images. In this step, you will create a CNN that classifies dog breeds. You must create your CNN from scratch (so, you can't use transfer learning yet!), and you must attain a test accuracy of at least 10%. In Step 4 of this notebook, you will have the opportunity to use transfer learning to create a CNN that attains greatly improved accuracy.

We mention that the task of assigning breed to dogs from images is considered exceptionally challenging. To see why, consider that even a human would have trouble distinguishing between a Brittany and a Welsh Springer Spaniel.

Brittany Welsh Springer Spaniel

It is not difficult to find other dog breed pairs with minimal inter-class variation (for instance, Curly-Coated Retrievers and American Water Spaniels).

Curly-Coated Retriever American Water Spaniel

Likewise, recall that labradors come in yellow, chocolate, and black. Your vision-based algorithm will have to conquer this high intra-class variation to determine how to classify all of these different shades as the same breed.

Yellow Labrador Chocolate Labrador Black Labrador

We also mention that random chance presents an exceptionally low bar: setting aside the fact that the classes are slightly imabalanced, a random guess will provide a correct answer roughly 1 in 133 times, which corresponds to an accuracy of less than 1%.

Remember that the practice is far ahead of the theory in deep learning. Experiment with many different architectures, and trust your intuition. And, of course, have fun!

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dog_images/train, dog_images/valid, and dog_images/test, respectively). You may find this documentation on custom datasets to be a useful resource. If you are interested in augmenting your training and/or validation data, check out the wide variety of transforms!

In [13]:
import os
from torchvision import datasets

### TODO: Write data loaders for training, validation, and test sets
## Specify appropriate transforms, and batch_sizes

img_dir = '/data/dog_images'
batch_size = 30
num_workers = 0

train_transform = transforms.Compose([transforms.Resize(255),
                                      transforms.RandomResizedCrop(224),
                                      transforms.RandomHorizontalFlip(),
                                      transforms.RandomRotation(10),
                                      transforms.ToTensor(),
                                      transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                            std=[0.229, 0.224, 0.225])])

validation_transform = transforms.Compose([transforms.Resize(255),
                                           transforms.CenterCrop(224),
                                           transforms.ToTensor(),
                                           transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                                 std=[0.229, 0.224, 0.225])])

test_transform = transforms.Compose([transforms.Resize(255),
                                           transforms.CenterCrop(224),
                                           transforms.ToTensor(),
                                           transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                                std=[0.229, 0.224, 0.225])])

train_data = datasets.ImageFolder(img_dir + '/train', transform = train_transform)
validation_data = datasets.ImageFolder(img_dir + '/valid', transform = validation_transform)
test_data = datasets.ImageFolder(img_dir + '/test', transform = test_transform)

train_loader = torch.utils.data.DataLoader(train_data, batch_size = batch_size, shuffle = True, num_workers = num_workers)
validation_loader = torch.utils.data.DataLoader(validation_data, batch_size = batch_size, shuffle = True, num_workers = num_workers)
test_loader = torch.utils.data.DataLoader(test_data, batch_size = batch_size, shuffle = False, num_workers = num_workers)

loaders_scratch = { 'train': train_loader,
                    'valid': validation_loader,
                    'test': test_loader }
In [14]:
print(test_loader)
<torch.utils.data.dataloader.DataLoader object at 0x7fee9d3ea978>

Question 3: Describe your chosen procedure for preprocessing the data.

  • How does your code resize the images (by cropping, stretching, etc)? What size did you pick for the input tensor, and why?
  • Did you decide to augment the dataset? If so, how (through translations, flips, rotations, etc)? If not, why not?

Answer:

  • I resized the images using transforms.Resize() with size = 226, and transforms.CenterCrop() with size = 224.
  • Yes I did, I augmented the training images by randomly fippling (transforms.RandomHorizontalFlip()) and resizing (transforms.RandomResizedCrop(224)) them , which might prevets overfitting since it gives randomness to the data. I didn't do any aumentation in both validation and testing sets.

(IMPLEMENTATION) Model Architecture

Create a CNN to classify dog breed. Use the template in the code cell below.

In [15]:
import torch.nn as nn
import torch.nn.functional as F

# define the CNN architecture
class Net(nn.Module):
    ### TODO: choose an architecture, and complete the class
    def __init__(self):
        super(Net, self).__init__()
        ## Define layers of a CNN
        self.conv1 = nn.Conv2d(3, 16, 3, padding = 1)
        self.conv2 = nn.Conv2d(16, 32, 3, padding = 1)
        self.conv3 = nn.Conv2d(32, 64, 3, padding = 1)
        self.conv4 = nn.Conv2d(64, 128, 3, padding = 1)
        self.conv5 = nn.Conv2d(128, 256, 3, padding = 1)
        
        # Pooling layers
        self.pool = nn.MaxPool2d(2,2)
        
        # Fully-connected layer
        self.fc1 = nn.Linear(256*7*7, 1500)
        self.fc2 = nn.Linear(1500, 500)
        self.fc3 = nn.Linear(500, 133)
        
        # Dropout
        self.dropout = nn.Dropout(0.25)
        
    def forward(self, x):
        ## Define forward behavior
        
        # First conve layer
        x = self.pool(F.relu(self.conv1(x)))
        # Second conve layer
        x = self.pool(F.relu(self.conv2(x)))
        # Third conve layer 
        x = self.pool(F.relu(self.conv3(x)))
        # Fourth conve layer 
        x = self.pool(F.relu(self.conv4(x)))
        # Fifth conve layer 
        x = self.pool(F.relu(self.conv5(x)))
        
        # flatten image input
        x = x.view(-1, 256*7*7)
        
        # Fully-conncted layer
        x = F.relu(self.fc1(x))
        x = self.dropout(x)
        x = F.relu(self.fc2(x))
        x = self.dropout(x)
        x = self.fc3(x)
        
        return x

#-#-# You do NOT have to modify the code below this line. #-#-#

# instantiate the CNN
model_scratch = Net()

# move tensors to GPU if CUDA is available
if use_cuda:
    model_scratch.cuda()
In [16]:
model_scratch
Out[16]:
Net(
  (conv1): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (conv2): Conv2d(16, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (conv3): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (conv4): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (conv5): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (fc1): Linear(in_features=12544, out_features=1500, bias=True)
  (fc2): Linear(in_features=1500, out_features=500, bias=True)
  (fc3): Linear(in_features=500, out_features=133, bias=True)
  (dropout): Dropout(p=0.25)
)

Question 4: Outline the steps you took to get to your final CNN architecture and your reasoning at each step.

Answer: In building this architecture, I followed same steps that I applied in CIFAR-10 excercise since it gave me a good result.

The architecture consists of 5 conve layers, and in each one I've applied kernel_size of 3, and padding of 1, which might prevent loosing information from images' corners. The output depth 'out_channel' starts with 16 and increased by (*2). After each conve layer, I've applied max pooling layer with kernal size and stride of 2, which downsize the image output by factor of 30 and final depth of 128.

I've finally applied 3 connceted layers in order to produce a final size of 133 (breeds' classes) and a dropout of 0.25 to prevent overfitting.

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_scratch, and the optimizer as optimizer_scratch below.

In [18]:
import torch.optim as optim

### TODO: select loss function
criterion_scratch = nn.CrossEntropyLoss()

### TODO: select optimizer
optimizer_scratch = optim.SGD(model_scratch.parameters(), lr = 0.025)

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_scratch.pt'.

In [19]:
# the following import is required for training to be robust to truncated images
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True

def train(n_epochs, loaders, model, optimizer, criterion, use_cuda, save_path):
    """returns trained model"""
    # initialize tracker for minimum validation loss
    valid_loss_min = np.Inf 
    
    for epoch in range(1, n_epochs+1):
        # initialize variables to monitor training and validation loss
        train_loss = 0.0
        valid_loss = 0.0
        
        ###################
        # train the model #
        ###################
        model.train()
        for batch_idx, (data, target) in enumerate(loaders['train']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            ## find the loss and update the model parameters accordingly
            ## record the average training loss, using something like
            ## train_loss = train_loss + ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            
            optimizer.zero_grad()
            
            # apply forward-pass
            output = model(data)
            
            # caluclate the loss
            loss = criterion(output, target)
            
            # apply backward-pass
            loss.backward()
            
            # optimization step (parameter update)
            optimizer.step()
            
            # update trainig loss
            train_loss = train_loss + ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            
        ######################    
        # validate the model #
        ######################
        model.eval()
        for batch_idx, (data, target) in enumerate(loaders['valid']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            ## update the average validation loss
            
            # apply forward-pass
            with torch.no_grad():
                output = model(data)
            
            # calculate the loss 
            loss = criterion(output, target)
            
            # update valudation loss
            valid_loss = valid_loss + ((1 / (batch_idx + 1)) * (loss.data - valid_loss))
            
        
        # calculate average losses
        train_loss = train_loss/len(loaders['train'].dataset)
        valid_loss = valid_loss/len(loaders['valid'].dataset)
            
        # print training/validation statistics 
        print('Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}'.format(
            epoch, 
            train_loss,
            valid_loss
            ))
        
        ## TODO: save the model if validation loss has decreased
        if valid_loss <= valid_loss_min:
            print('Validation loss decreased ({:.6f} --> {:.6f}).  Saving model ...'.format(
            valid_loss_min,
            valid_loss))
            torch.save(model.state_dict(), save_path)
            valid_loss_min = valid_loss
    # return trained model
    return model
In [20]:
# train the model
model_scratch = train(40, loaders_scratch, model_scratch, optimizer_scratch, 
                      criterion_scratch, use_cuda, 'model_scratch.pt')
Epoch: 1 	Training Loss: 0.000732 	Validation Loss: 0.005854
Validation loss decreased (inf --> 0.005854).  Saving model ...
Epoch: 2 	Training Loss: 0.000732 	Validation Loss: 0.005851
Validation loss decreased (0.005854 --> 0.005851).  Saving model ...
Epoch: 3 	Training Loss: 0.000731 	Validation Loss: 0.005848
Validation loss decreased (0.005851 --> 0.005848).  Saving model ...
Epoch: 4 	Training Loss: 0.000731 	Validation Loss: 0.005844
Validation loss decreased (0.005848 --> 0.005844).  Saving model ...
Epoch: 5 	Training Loss: 0.000730 	Validation Loss: 0.005838
Validation loss decreased (0.005844 --> 0.005838).  Saving model ...
Epoch: 6 	Training Loss: 0.000729 	Validation Loss: 0.005825
Validation loss decreased (0.005838 --> 0.005825).  Saving model ...
Epoch: 7 	Training Loss: 0.000728 	Validation Loss: 0.005816
Validation loss decreased (0.005825 --> 0.005816).  Saving model ...
Epoch: 8 	Training Loss: 0.000726 	Validation Loss: 0.005786
Validation loss decreased (0.005816 --> 0.005786).  Saving model ...
Epoch: 9 	Training Loss: 0.000722 	Validation Loss: 0.005733
Validation loss decreased (0.005786 --> 0.005733).  Saving model ...
Epoch: 10 	Training Loss: 0.000719 	Validation Loss: 0.005704
Validation loss decreased (0.005733 --> 0.005704).  Saving model ...
Epoch: 11 	Training Loss: 0.000715 	Validation Loss: 0.005631
Validation loss decreased (0.005704 --> 0.005631).  Saving model ...
Epoch: 12 	Training Loss: 0.000711 	Validation Loss: 0.005614
Validation loss decreased (0.005631 --> 0.005614).  Saving model ...
Epoch: 13 	Training Loss: 0.000703 	Validation Loss: 0.005535
Validation loss decreased (0.005614 --> 0.005535).  Saving model ...
Epoch: 14 	Training Loss: 0.000701 	Validation Loss: 0.005497
Validation loss decreased (0.005535 --> 0.005497).  Saving model ...
Epoch: 15 	Training Loss: 0.000697 	Validation Loss: 0.005491
Validation loss decreased (0.005497 --> 0.005491).  Saving model ...
Epoch: 16 	Training Loss: 0.000695 	Validation Loss: 0.005543
Epoch: 17 	Training Loss: 0.000691 	Validation Loss: 0.005390
Validation loss decreased (0.005491 --> 0.005390).  Saving model ...
Epoch: 18 	Training Loss: 0.000688 	Validation Loss: 0.005356
Validation loss decreased (0.005390 --> 0.005356).  Saving model ...
Epoch: 19 	Training Loss: 0.000684 	Validation Loss: 0.005321
Validation loss decreased (0.005356 --> 0.005321).  Saving model ...
Epoch: 20 	Training Loss: 0.000679 	Validation Loss: 0.005254
Validation loss decreased (0.005321 --> 0.005254).  Saving model ...
Epoch: 21 	Training Loss: 0.000675 	Validation Loss: 0.005218
Validation loss decreased (0.005254 --> 0.005218).  Saving model ...
Epoch: 22 	Training Loss: 0.000668 	Validation Loss: 0.005128
Validation loss decreased (0.005218 --> 0.005128).  Saving model ...
Epoch: 23 	Training Loss: 0.000665 	Validation Loss: 0.005110
Validation loss decreased (0.005128 --> 0.005110).  Saving model ...
Epoch: 24 	Training Loss: 0.000659 	Validation Loss: 0.005082
Validation loss decreased (0.005110 --> 0.005082).  Saving model ...
Epoch: 25 	Training Loss: 0.000656 	Validation Loss: 0.005020
Validation loss decreased (0.005082 --> 0.005020).  Saving model ...
Epoch: 26 	Training Loss: 0.000651 	Validation Loss: 0.004979
Validation loss decreased (0.005020 --> 0.004979).  Saving model ...
Epoch: 27 	Training Loss: 0.000649 	Validation Loss: 0.004982
Epoch: 28 	Training Loss: 0.000643 	Validation Loss: 0.004972
Validation loss decreased (0.004979 --> 0.004972).  Saving model ...
Epoch: 29 	Training Loss: 0.000639 	Validation Loss: 0.004916
Validation loss decreased (0.004972 --> 0.004916).  Saving model ...
Epoch: 30 	Training Loss: 0.000636 	Validation Loss: 0.004964
Epoch: 31 	Training Loss: 0.000632 	Validation Loss: 0.004837
Validation loss decreased (0.004916 --> 0.004837).  Saving model ...
Epoch: 32 	Training Loss: 0.000627 	Validation Loss: 0.004797
Validation loss decreased (0.004837 --> 0.004797).  Saving model ...
Epoch: 33 	Training Loss: 0.000623 	Validation Loss: 0.005014
Epoch: 34 	Training Loss: 0.000622 	Validation Loss: 0.004904
Epoch: 35 	Training Loss: 0.000616 	Validation Loss: 0.004654
Validation loss decreased (0.004797 --> 0.004654).  Saving model ...
Epoch: 36 	Training Loss: 0.000612 	Validation Loss: 0.004760
Epoch: 37 	Training Loss: 0.000610 	Validation Loss: 0.004696
Epoch: 38 	Training Loss: 0.000605 	Validation Loss: 0.004858
Epoch: 39 	Training Loss: 0.000599 	Validation Loss: 0.004608
Validation loss decreased (0.004654 --> 0.004608).  Saving model ...
Epoch: 40 	Training Loss: 0.000598 	Validation Loss: 0.004649
In [21]:
# load the model that got the best validation accuracy
model_scratch.load_state_dict(torch.load('model_scratch.pt'))

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 10%.

In [22]:
def test(loaders, model, criterion, use_cuda):

    # monitor test loss and accuracy
    test_loss = 0.
    correct = 0.
    total = 0.

    model.eval()
    for batch_idx, (data, target) in enumerate(loaders['test']):
        # move to GPU
        if use_cuda:
            data, target = data.cuda(), target.cuda()
        # forward pass: compute predicted outputs by passing inputs to the model
        output = model(data)
        # calculate the loss
        loss = criterion(output, target)
        # update average test loss 
        test_loss = test_loss + ((1 / (batch_idx + 1)) * (loss.data - test_loss))
        # convert output probabilities to predicted class
        pred = output.data.max(1, keepdim=True)[1]
        # compare predictions to true label
        correct += np.sum(np.squeeze(pred.eq(target.data.view_as(pred))).cpu().numpy())
        total += data.size(0)
            
    print('Test Loss: {:.6f}\n'.format(test_loss))

    print('\nTest Accuracy: %2d%% (%2d/%2d)' % (
        100. * correct / total, correct, total))
In [23]:
# call test function    
test(loaders_scratch, model_scratch, criterion_scratch, use_cuda)
Test Loss: 3.852525


Test Accuracy: 11% (94/836)

Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)

You will now use transfer learning to create a CNN that can identify dog breed from images. Your CNN must attain at least 60% accuracy on the test set.

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dogImages/train, dogImages/valid, and dogImages/test, respectively).

If you like, you are welcome to use the same data loaders from the previous step, when you created a CNN from scratch.

In [24]:
## TODO: Specify data loaders

img_dir = '/data/dog_images'
batch_size = 30
num_workers = 0

train_transform = transforms.Compose([transforms.Resize(255),
                                      transforms.RandomResizedCrop(224),
                                      transforms.RandomRotation(15),
                                      transforms.ToTensor(),
                                      transforms.Normalize([0.5, 0.5, 0.5],
                                                           [0.5, 0.5, 0.5])])

validation_transform = transforms.Compose([transforms.Resize(255),
                                           transforms.CenterCrop(224),
                                           transforms.ToTensor(),
                                           transforms.Normalize([0.5, 0.5, 0.5],
                                                                [0.5, 0.5, 0.5])])

test_transform = transforms.Compose([transforms.Resize(255),
                                           transforms.CenterCrop(224),
                                           transforms.ToTensor(),
                                           transforms.Normalize([0.5, 0.5, 0.5],
                                                                [0.5, 0.5, 0.5])])

train_data = datasets.ImageFolder(img_dir + '/train', transform = train_transform)
validation_data = datasets.ImageFolder(img_dir + '/valid', transform = validation_transform)
test_data = datasets.ImageFolder(img_dir + '/test', transform = test_transform)

train_loader = torch.utils.data.DataLoader(train_data, batch_size = batch_size, shuffle = True, num_workers = num_workers)
validation_loader = torch.utils.data.DataLoader(validation_data, batch_size = batch_size, shuffle = True, num_workers = num_workers)
test_loader = torch.utils.data.DataLoader(test_data, batch_size = batch_size, shuffle = False, num_workers = num_workers)

loaders_transfer = { 'train': train_loader,
                    'valid': validation_loader,
                    'test': test_loader }

(IMPLEMENTATION) Model Architecture

Use transfer learning to create a CNN to classify dog breed. Use the code cell below, and save your initialized model as the variable model_transfer.

In [25]:
import torchvision.models as models
import torch.nn as nn

## TODO: Specify model architecture 

# Load the pretrained model 
model_transfer = models.resnet18(pretrained=True)

print(model_transfer)
Downloading: "https://download.pytorch.org/models/resnet18-5c106cde.pth" to /root/.torch/models/resnet18-5c106cde.pth
100%|██████████| 46827520/46827520 [00:00<00:00, 88251153.28it/s]
ResNet(
  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU(inplace)
  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AvgPool2d(kernel_size=7, stride=1, padding=0)
  (fc): Linear(in_features=512, out_features=1000, bias=True)
)
In [27]:
# number of classes 
classes = len(train_data.classes)

# redefine the last layer 
inputs = model_transfer.fc.in_features
model_transfer.fc = nn.Linear(inputs, classes, bias = True)

# Freeze training for the features layers
for param in model_transfer.parameters():
    param.requires_grad = False
    
# Unfreeze training for fully connected layers   
for param in model_transfer.fc.parameters():
    param.requires_grad = True

print(model_transfer)
ResNet(
  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU(inplace)
  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AvgPool2d(kernel_size=7, stride=1, padding=0)
  (fc): Linear(in_features=512, out_features=133, bias=True)
)
In [28]:
if use_cuda:
    model_transfer = model_transfer.cuda()

Question 5: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. Describe why you think the architecture is suitable for the current problem.

Answer:

I've read some articles that explain and compare different architectures [1][2], each architecture has its pros and cons. Thus, I've chosen ResNet in order to do this step. ResNet is considered as one of the good model that has an excellent generalization performance with fewer error rates on recognition tasks such as this.

I've replaced the final connected layer (fc) of the architecure with a connected layer with output of 133 (dogs breeds)

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_transfer, and the optimizer as optimizer_transfer below.

In [30]:
criterion_transfer = nn.CrossEntropyLoss()
optimizer_transfer = optim.SGD(model_transfer.fc.parameters(), lr = 0.007)

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_transfer.pt'.

In [31]:
# train the model
model_transfer = train(40, loaders_transfer, model_transfer, optimizer_transfer, criterion_transfer, use_cuda, 'model_transfer.pt')
Epoch: 1 	Training Loss: 0.000698 	Validation Loss: 0.004879
Validation loss decreased (inf --> 0.004879).  Saving model ...
Epoch: 2 	Training Loss: 0.000601 	Validation Loss: 0.003946
Validation loss decreased (0.004879 --> 0.003946).  Saving model ...
Epoch: 3 	Training Loss: 0.000524 	Validation Loss: 0.003282
Validation loss decreased (0.003946 --> 0.003282).  Saving model ...
Epoch: 4 	Training Loss: 0.000470 	Validation Loss: 0.002775
Validation loss decreased (0.003282 --> 0.002775).  Saving model ...
Epoch: 5 	Training Loss: 0.000425 	Validation Loss: 0.002425
Validation loss decreased (0.002775 --> 0.002425).  Saving model ...
Epoch: 6 	Training Loss: 0.000388 	Validation Loss: 0.002133
Validation loss decreased (0.002425 --> 0.002133).  Saving model ...
Epoch: 7 	Training Loss: 0.000364 	Validation Loss: 0.001953
Validation loss decreased (0.002133 --> 0.001953).  Saving model ...
Epoch: 8 	Training Loss: 0.000341 	Validation Loss: 0.001766
Validation loss decreased (0.001953 --> 0.001766).  Saving model ...
Epoch: 9 	Training Loss: 0.000322 	Validation Loss: 0.001659
Validation loss decreased (0.001766 --> 0.001659).  Saving model ...
Epoch: 10 	Training Loss: 0.000308 	Validation Loss: 0.001524
Validation loss decreased (0.001659 --> 0.001524).  Saving model ...
Epoch: 11 	Training Loss: 0.000296 	Validation Loss: 0.001430
Validation loss decreased (0.001524 --> 0.001430).  Saving model ...
Epoch: 12 	Training Loss: 0.000288 	Validation Loss: 0.001356
Validation loss decreased (0.001430 --> 0.001356).  Saving model ...
Epoch: 13 	Training Loss: 0.000272 	Validation Loss: 0.001293
Validation loss decreased (0.001356 --> 0.001293).  Saving model ...
Epoch: 14 	Training Loss: 0.000269 	Validation Loss: 0.001237
Validation loss decreased (0.001293 --> 0.001237).  Saving model ...
Epoch: 15 	Training Loss: 0.000258 	Validation Loss: 0.001219
Validation loss decreased (0.001237 --> 0.001219).  Saving model ...
Epoch: 16 	Training Loss: 0.000257 	Validation Loss: 0.001164
Validation loss decreased (0.001219 --> 0.001164).  Saving model ...
Epoch: 17 	Training Loss: 0.000245 	Validation Loss: 0.001110
Validation loss decreased (0.001164 --> 0.001110).  Saving model ...
Epoch: 18 	Training Loss: 0.000243 	Validation Loss: 0.001105
Validation loss decreased (0.001110 --> 0.001105).  Saving model ...
Epoch: 19 	Training Loss: 0.000239 	Validation Loss: 0.001077
Validation loss decreased (0.001105 --> 0.001077).  Saving model ...
Epoch: 20 	Training Loss: 0.000236 	Validation Loss: 0.001057
Validation loss decreased (0.001077 --> 0.001057).  Saving model ...
Epoch: 21 	Training Loss: 0.000231 	Validation Loss: 0.001008
Validation loss decreased (0.001057 --> 0.001008).  Saving model ...
Epoch: 22 	Training Loss: 0.000227 	Validation Loss: 0.000997
Validation loss decreased (0.001008 --> 0.000997).  Saving model ...
Epoch: 23 	Training Loss: 0.000225 	Validation Loss: 0.000974
Validation loss decreased (0.000997 --> 0.000974).  Saving model ...
Epoch: 24 	Training Loss: 0.000219 	Validation Loss: 0.000950
Validation loss decreased (0.000974 --> 0.000950).  Saving model ...
Epoch: 25 	Training Loss: 0.000219 	Validation Loss: 0.000929
Validation loss decreased (0.000950 --> 0.000929).  Saving model ...
Epoch: 26 	Training Loss: 0.000210 	Validation Loss: 0.000917
Validation loss decreased (0.000929 --> 0.000917).  Saving model ...
Epoch: 27 	Training Loss: 0.000215 	Validation Loss: 0.000921
Epoch: 28 	Training Loss: 0.000211 	Validation Loss: 0.000896
Validation loss decreased (0.000917 --> 0.000896).  Saving model ...
Epoch: 29 	Training Loss: 0.000212 	Validation Loss: 0.000892
Validation loss decreased (0.000896 --> 0.000892).  Saving model ...
Epoch: 30 	Training Loss: 0.000205 	Validation Loss: 0.000872
Validation loss decreased (0.000892 --> 0.000872).  Saving model ...
Epoch: 31 	Training Loss: 0.000206 	Validation Loss: 0.000849
Validation loss decreased (0.000872 --> 0.000849).  Saving model ...
Epoch: 32 	Training Loss: 0.000206 	Validation Loss: 0.000853
Epoch: 33 	Training Loss: 0.000200 	Validation Loss: 0.000845
Validation loss decreased (0.000849 --> 0.000845).  Saving model ...
Epoch: 34 	Training Loss: 0.000200 	Validation Loss: 0.000840
Validation loss decreased (0.000845 --> 0.000840).  Saving model ...
Epoch: 35 	Training Loss: 0.000199 	Validation Loss: 0.000848
Epoch: 36 	Training Loss: 0.000196 	Validation Loss: 0.000837
Validation loss decreased (0.000840 --> 0.000837).  Saving model ...
Epoch: 37 	Training Loss: 0.000197 	Validation Loss: 0.000829
Validation loss decreased (0.000837 --> 0.000829).  Saving model ...
Epoch: 38 	Training Loss: 0.000193 	Validation Loss: 0.000806
Validation loss decreased (0.000829 --> 0.000806).  Saving model ...
Epoch: 39 	Training Loss: 0.000193 	Validation Loss: 0.000793
Validation loss decreased (0.000806 --> 0.000793).  Saving model ...
Epoch: 40 	Training Loss: 0.000192 	Validation Loss: 0.000807
In [32]:
# load the model that got the best validation accuracy (uncomment the line below)
model_transfer.load_state_dict(torch.load('model_transfer.pt'))

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 60%.

In [33]:
test(loaders_transfer, model_transfer, criterion_transfer, use_cuda)
Test Loss: 0.674154


Test Accuracy: 82% (687/836)

(IMPLEMENTATION) Predict Dog Breed with the Model

Write a function that takes an image path as input and returns the dog breed (Affenpinscher, Afghan hound, etc) that is predicted by your model.

In [63]:
### TODO: Write a function that takes a path to an image as input
### and returns the dog breed that is predicted by the model.

# list of class names by index, i.e. a name can be accessed like class_names[0]
class_names = [item[4:].replace("_", " ") for item in train_data.classes]

def predict_breed_transfer(img_path):
    
    transform = transforms.Compose([transforms.Resize(255),
                                      transforms.CenterCrop(224),
                                      transforms.ToTensor(),
                                      transforms.Normalize([0.5, 0.5, 0.5],
                                                           [0.5, 0.5, 0.5])])
    
    
    # load the image and return the predicted breed
    
    # load the image
    img = Image.open(img_path).convert('RGB')
    
    # apply preprocessing
    img = transform(img)
    img = img.unsqueeze(0)
    
    if use_cuda:
        img = img.cuda()
        
    model_transfer.eval()
    
    output = model_transfer(img)
    pre = torch.argmax(output).item()
    
    return class_names[pre]

Step 5: Write your Algorithm

Write an algorithm that accepts a file path to an image and first determines whether the image contains a human, dog, or neither. Then,

  • if a dog is detected in the image, return the predicted breed.
  • if a human is detected in the image, return the resembling dog breed.
  • if neither is detected in the image, provide output that indicates an error.

You are welcome to write your own functions for detecting humans and dogs in images, but feel free to use the face_detector and human_detector functions developed above. You are required to use your CNN from Step 4 to predict dog breed.

Some sample output for our algorithm is provided below, but feel free to design your own user experience!

Sample Human Output

(IMPLEMENTATION) Write your Algorithm

In [69]:
### TODO: Write your algorithm.
### Feel free to use as many code cells as needed.

def run_app(img_path):
    ## handle cases for a human face, dog, and neither
    if(face_detector(img_path)):
        print('Hello, human!')
        plt.imshow(Image.open(img_path))
        plt.show()
        print('You look like a ... {0}'.format(predict_breed_transfer(img_path)))
    elif(dog_detector(img_path)):
        print('Hello, dog!')
        plt.imshow(Image.open(img_path))
        plt.show()
        print('You look like a ... {0}'.format(predict_breed_transfer(img_path)))
    else:
        plt.imshow(Image.open(img_path))
        plt.show()
        print('Hello! sorry, I could not recognize what you are')
        
    print('--------------------')

Step 6: Test Your Algorithm

In this section, you will take your new algorithm for a spin! What kind of dog does the algorithm think that you look like? If you have a dog, does it predict your dog's breed accurately? If you have a cat, does it mistakenly think that your cat is a dog?

(IMPLEMENTATION) Test Your Algorithm on Sample Images!

Test your algorithm at least six images on your computer. Feel free to use any images you like. Use at least two human and two dog images.

Question 6: Is the output better than you expected :) ? Or worse :( ? Provide at least three possible points of improvement for your algorithm.

Answer: (Three possible points for improvement)

I didn't expecte the output to be good, but not that bad too :(. I've re-run the model multiple times with many changes (adding more layers, increasing and decreasing learning rate and epochs, use different augmentation parameters, etc) in order to get a good result; however, the accuracy always ranges between 11%-16%.

I believe that if we want to get higher results, we can:

  • add more samples per category, which will improve the training. Especially since the data is considered a bit challenging (different breeds look similar)
  • Add more image augmentations to increase the randomness of the dataset
  • Try and improve different pretrained models
  • Try different hyper parameter tunings
In [61]:
## TODO: Execute your algorithm from Step 6 on
## at least 6 images on your computer.
## Feel free to use as many code cells as needed.

## suggested code, below
for file in np.hstack((human_files[:3], dog_files[:3])):
    run_app(file)
Hello, human!
You look like a ... Chihuahua
--------------------
Hello, human!
You look like a ... Dalmatian
--------------------
Hello, human!
You look like a ... American water spaniel
--------------------
Hello, dog!
You look like a ... Mastiff
--------------------
Hello, dog!
You look like a ... Mastiff
--------------------
Hello, dog!
You look like a ... Bullmastiff
--------------------
In [70]:
images = np.array(glob("RandomImages/*"))
for path in images:
    run_app(path)    
Hello! sorry, I could not recognize what you are
--------------------
Hello, dog!
You look like a ... Alaskan malamute
--------------------
Hello, human!
You look like a ... Dachshund
--------------------
Hello! sorry, I could not recognize what you are
--------------------
Hello, human!
You look like a ... Dalmatian
--------------------
Hello, dog!
You look like a ... Irish setter
--------------------